Frédéric Ratle 

Welcome to my personal web page! 


Hi! Somehow you ended up on my web page. It is now mostly a repository for my papers, but I do update it from time to time  :-)

I can easily be reached via Gmail if you put my first name followed by a dot and my last name.  



I completed my B.Eng. and M.Sc. at the Ecole Polytechnique de Montréal (Canada) and my PhD at the IGAR institute of the University of Lausanne (Switzerland).  My thesis dealt with unsupervised machine learning, more specifically adaptive and efficient methods for the clustering of large databases.  I now work as a scientist at Nuance Communications.  


Conference papers

F. Ratle, J. Weston, M. L. Miller. Large-scale clustering through functional embedding. European Conference on Machine Learning (ECML 2008), 2008.

J. Weston, F. Ratle, R. Collobert. Deep learning via semi-supervised embedding. Int. Conference on Machine Learning (ICML 2008), 2008. 

F. Ratle, A.L. Terrettaz-Zufferey, M. Kanevski, P. Esseiva, O. Ribaux. A comparison of one-class classifiers for novelty detection in forensic case data. Int. Conf. on Intelligent Data Engineering and Automated Learning, Springer, 2007. 

F. Ratle, D. Tuia. Ensemble methods for environmental data modelling with support vector regression. European Colloquium on Theoretical and Quantitative Geography, 2007.

F. Ratle, A.L. Terrettaz, M. Kanevski, P. Esseiva, O. Ribaux.  Learning manifolds in forensic data, International Conference on Artificial Neural Networks (ICANN), Springer, 2006.

F. Ratle, A.L. Terrettaz, M. Kanevski, P. Esseiva, O. Ribaux. Pattern analysis in illicit heroin seizures: a novel application of machine learning algorithms. Verleysen et al. (eds): European Symposium on Artificial Neural Networks (ESANN), d-side publi., 2006.

F. Ratle, B. Lecarpentier, R. Labib, F. Trochu. Multi-objective optimization of a composite material spring design using an evolutionary algorithm. Yao et al. (eds) : Parallel Problem Solving from Nature (PPSN VIII) 803-811, Springer, 2004. 


Journal papers

F. Ratle, G. Camps-Valls, J. Weston. Semi-supervised neural networks for hyperspectral image classification. IEEE Transactions on Geoscience and Remote Sensing, 48(5), 2010. 

D. Tuia, F. Ratle, F. Pacifici, M. Kanevski, and W. J. Emery.  Active learning methods for remote sensing image classification. IEEE Transactions on Geoscience and Remote Sensing 47(7):2218-2232, 2009.

F. Ratle, C, Gagné, A.L. Terrettaz-Zufferey, M. Kanevski, P. Esseiva, O. Ribaux. Advanced clustering methods for mining chemical databases in forensic science. Journal of Chemometrics and Intelligent Laboratory Systems 90(2): 123-131, 2008.

D. Tuia, F. Ratle, R. Lasaponara, L. Telesca, M. Kanevski. Scan statistics analysis of forest fire clusters, Journal of Communications in Nonlinear Sciences and Numerical Simulations 13(8): 1689-1694, 2007.

A.L. Terrettaz-Zufferey, F. Ratle, M. Kanevski, P. Esseiva, O. Ribaux. Pattern detection in forensic case data using graph theory: Application to heroin cutting agents, Forensic Science International, Volume 167, Issues 2-3, Pages 242-246, 2007.


Book chapters

J. Weston, F. Ratle, H. Mobahi, R. Collobert. Deep Learning via Semi-Supervised Embedding. In: Neural Networks Tricks of the Trade, Reloaded, Springer LNCS, 2012.

F. Ratle, A. Pozdnoukhov, V. Demyanov, V. Timonin, E. Savelieva. Spatial data analysis and mapping using machine learning algorithms.  In Kanevski (ed.):  Advanced Mapping of Environmental Data - Geostatistics, Machine Learning and Bayesian Maximum Entropy, ISTE, 2008.


Talks and posters

F. Ratle. Spectral clustering and one-class classification: new tools for forensic drug intelligence. NATO Advanced Study Institute on Mining Massive Datasets for Security, Gazzada (Italy), 2007.

F. Ratle. Space-time cluster detection in crime data with scan statistics. Spatial Econometrics Association Conference, Cambridge (UK), 2007.


Unpublished stuff 

Ratle, F. Optimisation de la conception et de la fabrication des matériaux composites par des algorithmes d'évolution et d'apprentissage. Master's thesis, Ecole Polytechnique de Montréal, 2005.



Probably not-so-useful spectral clustering and laplacian eigenmaps matlab code here

(Note: some bits have been borrowed here and there; original authors are credited where appropriate.)



Friends, colleagues and recent co-authors

Gustavo Camps-Valls        University of Valencia, Spain

Christian Gagné               Université Laval, Québec , Canada

Jason Weston                 Google Research NY, USA

Devis Tuia                      EPFL, Switzerland